Documente Academic
Documente Profesional
Documente Cultură
sad0616Shla3p.indd 52
4/20/16 6:21 PM
THE
TRUTH
ABOUT
SELFDRIVING
CARS
T R A N S P O R TAT I O N
*****************************************
*****************************************
*****************************************
* * * * * * * * * * * *A*I * * * * * * * * * * * * * * * * * * * * * * * * * * * *
* * * * * * * * *S*P*E*C*I*A*L* * * * * * * * * * * * * * * * * * * * * * * * * *
*****************************************
* * * * * * * * * R* E* P* O
*****************************
* * * * * * * * * * * * * *R*T* * * * * * * * * * * * * * * * * * * * * * * * * *
*****************************************
*****************************************
*****************************************
sad0616Shla3p.indd 53
4/20/16 6:21 PM
DRIVING IS A MUCH MORE COMPLEX ACTIVITY than most people appreciate. It involves a broad range of skills and actions, some of
which are easier to automate than others. Maintaining speed on
an open road is simple, which is why conventional cruise-control systems have been doing it automatically for decades. As
technology has advanced, engineers have been able to automate
additional driving subtasks. Widely available adaptive cruisecontrol systems now maintain proper speed and spacing behind
other vehicles. Lane-keeping systems, such as those in new
models from Mercedes-Benz and Infiniti, use cameras, sensors
and steering control to keep a vehicle centered in its lane. Cars
are pretty smart these days. Yet it is an enormous leap from such
systems to fully automated driving.
A five-level taxonomy defined by SAE International (formerly
IN BRIEF
Automatic parking valets, low-speed campus shuttles, closely spaced platoons of heavy trucks and automatic freeway-control systems for use in dedicated
lanes are all feasible and perhaps inevitable.
sad0616Shla3p.indd 54
4/20/16 5:30 PM
*****************************
TA XO N O M Y
Who steers,
accelerates
and
decelerates
No
Automation
Driver
Assistance
Partial
Automation
Conditional
Automation
High
Automation
Full
Automation
The combination
of automatic speed
and steering controlfor example,
cruise control and
lane keeping.
Automated systems
that do everythingno human
backup required
but only in limited
circumstances.
Human driver
Human driver
and system
System
System
System
System
Human driver
Human driver
Human driver
System
System
System
Human driver
Human driver
Human driver
Human driver
System
System
None
Some driving
modes
Some driving
modes
Some driving
modes
Some driving
modes
Who monitors
the driving
environment
Who takes
control when
something
goes wrong
How much
driving,
overall, is
assisted or
automated
sad0616Shla3p.indd 55
4/20/16 5:30 PM
NEXT YEAR Volvo Cars will field-test 100 vehicles equipped with
systems that automate driving on special stretches of freeway (1 and
2). Volvos have also been used in European road-train tests (3).
1
Despite the popular perception, h
uman drivers are re
markably capable of avoiding serious crashes. Based
on the total U.S. traffic safety statistics for 2011, fatal crashes
occurred about once for every 3.3million hours of driving; crashes that resulted in injury happened approximately once for every
64,000 hours of driving. These numbers set an important safety
target for automated driving systems, which should, at minimum, be no less safe than human drivers. Reaching this level of
reliability will require vastly more development than automation
enthusiasts want to admit.
Think about how often your laptop freezes up. If that software
were responsible for driving a car, the blue screen of death
would become more than a figure of speech. A delayed software
response of as little as one tenth of a second is likely to be hazardous in traffic. Software for automated driving must therefore be
designed and developed to dramatically different standards from
anything currently found in consumer devices.
Achieving these standards will be profoundly difficult and re
quire basic breakthroughs in software engineering and signal processing. Engineers need new methods for designing software that
can be proved correct and safe even in complex and rapidly changing conditions. Formal methods for analyzing every possible failure mode for a piece of code before it is written existthink of
them as mathematical proofs for computer programsbut only
for very simple applications. Scientists are only beginning to think
about how to scale up these kinds of tests to validate the incredibly complex code required to control a fully automated vehicle.
Once that code has been written, software engineers will need
new methods for debugging and verifying it. Existing methods
are too cumbersome and costly for the job. To put this in perspective, consider that half of the cost of a new commercial or military
aircraft goes toward software verification and validation. The
software on aircraft is actually much less c omplex than what will
be needed for automated road vehicles. An engineer can design
an aircraft autopilot system knowing that it will rarely, if ever,
have to deal with more than one or two other aircraft in its vicinity. It does not need to know the velocity and location of those aircraft with incredible precision, because they are far enough apart
that they have time to act. Decisions must be made on the order
of tens of seconds. An automated road vehicle will have to track
dozens of other vehicles and obstacles and make decisions within fractions of a second. The code required will be orders of
sad0616Shla3p.indd 56
4/20/16 5:30 PM
that were not identified) and extremely low false positives (be
nign objects that were misclassified, leading to inappropriate re
sponses from vehicles, such as swerving or hard braking).
Engineers cannot resort to the kind of brute-force redundan
cy used in commercial aircraft systems to achieve these goals be
cause an automated car is a consumer product: it must be af
fordable for the general public. Turning to artificial intelligence
is not an obvious solution, either. Some people have suggested
that machine-learning systems could enable automated driving
systems to study millions of hours of driving data and then learn
throughout the course of their life cycle. But machine learning
introduces its own problems because it is nondeterministic.
Two identical vehicles can roll off the assembly line, but after a
year of encountering different traffic situations, their automa
tion systems will behave very differently.
LEVEL-FOUR FUTURE
M O R E TO E X P L O R E
sad0616Shla3p.indd 57
4/20/16 5:30 PM